CONVERGENCE OF CHANGES IN THE EMPLOYMENT STRUCTURE IN THE EUROPEAN UNION COUNTRIES

Purpose: Multidimensional scaling is a method that allows the representation of complex observations in a space with fewer dimensions (usually on the plane) in such a way that the distance matrix counted on the plane is as similar as possible to the matrix counted in the original space. The paper presents a proposal to use this approach in the analysis of spatial temporal structures describing changes in the structure of employment in EU countries. The structures are described by shares, so no procedure is necessary to bring the shares to comparability due to the unit of measure. Separate multidimensional scaling is carried out for each unit of time. The purpose of the study is to try to separate groups of European Union countries similar due to their employment structures. Methodology: The paper uses one of the frequently used non-hierarchical methods - the k-means method. The study of multidimensional clustering of EU regions, which uses the k-means algorithm to identify similar areas are used in many fields among others to group regions or countries. Findings: In the EU countries, enterprises employing up to 9 people account for the largest percentage. On average, they accounted for 95.14% in 2015 and in 2020 the percentage of these enterprises increased slightly to 95.27%. The highest percentage of enterprises employing up to 9 people was in Slovakia, with 97.41% in 2020, while the lowest was in Germany, where the value of this indicator was 82.44%. Among the countries analysed, Germany and Luxembourg stood out in particular, where there was a relatively high number of companies employing between 50 and 249 people. Originality/value: The importance of the indicated research is extremely important from the point of view of EU countries. Therefore, the methods indicated in the article can also be used in other aspects of the economy.